Effects of Changing Monetary and Regulatory Policy on ...

Effects of Changing Monetary and RegulatoryPolicy on Money Markets∗

Elizabeth Klee, Zeynep Senyuz, and Emre YoldasFederal Reserve Board

The global financial crisis and the resulting policy responseled to substantial changes in U.S. dollar funding markets,which are crucial for the functioning of the financial systemand the transmission of monetary policy in the United States.We develop and test hypotheses on the effects of changing mon-etary and regulatory policy on key funding rates. We showthat the federal funds rate continued to provide an anchor forunsecured rates, albeit weaker, while its transmission to thesecured repo rate is hampered in the post-crisis period. TheFederal Reserve’s reverse repurchase facility led to strongerco-movement and reduced volatility of money market rates.The new regulations and the superabundant reserves envi-ronment affected rate dynamics on calendar days primarilythrough increased balance sheet costs.

JEL Codes: C32, E43, E52, G21, G28.

1. Introduction

During the global financial crisis of 2007 to 2009, and in the yearsfollowing, the Federal Reserve (Fed) injected massive amounts of

∗We would like to thank James Clouse, Jane Ihrig, Fulvio Pegoraro, and par-ticipants at the 2017 European Central Bank Money Markets Workshop, 2016International Finance and Banking Society Barcelona Conference, 2016 WesternEconomic Association International Conference, and UC Riverside EconomicsDepartment seminar for helpful comments. We thank Richard Sambasivam forhis assistance with the data set and Gurubala Kotta for her editorial suggestions.The views expressed in this paper are solely the responsibility of the authorsand should not be interpreted as reflecting the views of the Board of Governorsof the Federal Reserve System, or other members of its staff. Author e-mails:[email protected]; [email protected]; [email protected].

165

166 International Journal of Central Banking October 2019

liquidity into the financial system, kept its target policy rate nearzero, and introduced new tools to conduct monetary policy. Mean-while, a series of regulatory reforms that prompted financial institu-tions to reevaluate their risk-management practices were announcedand implemented. These changes to monetary policy implementationand the regulatory environment marked a new era for the financialsystem. In this paper we focus on the impact of these developmentson money markets.

Understanding the effects of changes to the Fed’s monetary pol-icy framework and financial regulations is important for a number ofreasons. To start, the Fed uses money markets to influence broaderfinancial conditions. Specifically, the first step of monetary policytransmission is for the Fed’s policy rate (the federal funds rate) toinfluence other overnight interest rates. For effective policy imple-mentation, other overnight rates should move closely with the fed-eral funds rate, that is, after controlling for risk factors and marketfrictions, rate differentials should be arbitraged away. Impeded mon-etary policy transmission could pose challenges to the central bankfor controlling interest rates. Therefore, it is important to identifyhow and to what extent pass-through from the federal funds rateto other money market interest rates has been affected since thecrisis.

In addition, some adjustments to the regulatory framework thatwere aimed at money markets also created new incentives for marketparticipants. Reliance of dealers and other financial intermediarieson short-term borrowing in money markets may have contributedin part to the financial crisis. Consequently, many regulationsfocused on reducing risks to institutions operating in money markets.Understanding the intended or unintended consequences of theseactions helps policymakers evaluate the costs and benefits of thesepolicies.

Finally, money markets represent a significant share of finan-cial intermediation. Therefore, quantifying the effects of changes inmonetary policy implementation and regulation on these marketsis important. To provide context, at the end of the first quarter of2018, there were roughly $4.5 trillion in money market instrumentsoutstanding, with federal funds and repurchase agreements (repos)alone representing $3.5 trillion. As a point of comparison, there were

Vol. 15 No. 4 Effects of Changing Monetary and Regulatory Policy 167

about $5.4 trillion outstanding in non-financial corporate bonds inthe same period.1

Against this backdrop, we document significant changes to theFed’s monetary policy implementation framework and to the reg-ulatory environment over the course of the financial crisis and thesubsequent effective lower bound (ELB) period. We then identifytheir effects on dynamics of key overnight funding rates. To do so,we estimate dynamic multivariate models for money market ratesover two sample periods: the pre-crisis period that runs from Janu-ary 2001 to 2007 and serves as a benchmark, and the ELB periodfrom December 2008 to August 2015, during which the aforemen-tioned policy changes took place. Our pre-crisis model incorporatesthe long-run relationship of the federal funds rate with the otherovernight rates in a multivariate framework. We assume time varia-tion in the volatilities and correlations of funding rates to allow forpotentially different dynamics around financial reporting days, whensome institutions contract their balance sheets.

Our results suggest that despite important changes in the marketstructure, the federal funds rate continued to provide an anchor forunsecured overnight rates. At the same time, the co-movement ofrates weakened significantly. In particular, transmission of the fed-eral funds rate to the repo rate was hampered. We also illustratesome differences in behavior for unsecured versus secured moneymarket rates. Specifically, we show that the new regulations sub-stantially altered unsecured rate dynamics on financial reportingdays by increasing balance sheet costs of financial intermediaries.Rates that represent unsecured wholesale funding costs for banksbecame significantly lower and more volatile on quarter-ends. Bycontrast, for secured rates, quarter-end effects weakened in the repomarket, reflecting lower dealer leverage in response to new regula-tions and reduced net repo financing. Separately, day-of-week effectson the federal funds rate mostly disappeared due to the abundanceof reserve balances post-crisis.

In addition to exploring differences between the pre-crisis andELB periods, we also take a close look at structural changes dur-ing the post-crisis period. We use September 2013 as a natural

1Refer to Federal Reserve Z.1 release, Financial Accounts, tables 209 and 213.


breakpoint in this sample, as two major developments took placearound that time. First, the Fed expanded its monetary policytoolkit with the introduction of the overnight reverse repurchase(ON RRP) facility. Second, a number of Basel III regulatory changeswere announced around that time. We document that the ON RRPfacility strengthened the link between the repo rates and unsecuredrates, and also contributed to better monetary policy transmission.Moreover, volatility of all rates dampened, with an especially notabledecline in the repo market. We also find that the tendency of foreignbanks to reduce their overnight borrowing on financial-reporting-related days, combined with the search by cash lenders for alterna-tive investment opportunities, exacerbated month-end and quarter-end effects on the federal funds rate and Eurodollar rates. The avail-ability of the ON RRP as a viable investment option on financialreporting days, when alternatives are limited, reduced the potentialfor sharp drops in the repo rate.

In related literature on money market dynamics, Afonso, Kovner,and Schoar (2011) analyze activity in the federal funds market dur-ing the global financial crisis, while Gorton and Metrick (2012) andCopeland, Martin, and Walker (2014) focus on the role of repo mar-kets during the crisis in the context of runs. Yoldas and Senyuz(2015) model the behavior of term money market rates and quantifyfinancial stress. Although the literature on monetary policy trans-mission to the economy is vast, there is relatively limited research onhow the target rate is transmitted to other overnight interest rates.Bech, Klee, and Stebunovs (2014) find evidence of deterioration ofthe pass-through from the federal funds rate to the repo rate duringthe financial crisis that seemed to persist, while Kroeger and Sarkar(2016) suggest that this pass-through improved with the ON RRPfacility.

Another strand of literature related to our work documents theeffects of calendar days on money market rates. Spindt and Hoffmeis-ter (1988), Griffiths and Winters (1995), Hamilton (1996), Carpenterand Demiralp (2006), and Judson and Klee (2010) show that thefederal funds rate exhibits calendar-day effects associated with themaintenance period as well as quarter-ends. More recently, Mun-yan (2015) and Anbil and Senyuz (2016) document the effects ofwindow-dressing activity on financial reporting dates in the repomarket.


The rest of the paper is organized as follows. In the next section,we provide background information on the mechanics of money mar-kets, review changes in the monetary policy implementation frame-work and relevant regulations, and establish hypotheses on theireffects. We describe the data set in section 3 and lay out the method-ological framework in section 4. We present and discuss the estima-tion results in section 5. We conclude in section 6.

2. Money Markets and the Changing Landscape

The response of the Fed to the global financial crisis of 2007–09 sig-nificantly altered the money market landscape in the United States.With successive rate cuts, the target federal funds rate was reducedfrom 5.25 percent in August 2007 to its effective lower bound (ELB)of 0 to 0.25 percent in December 2008. The federal funds rate, aswell as other overnight rates, remained at the ELB for the next sevenyears. Throughout the crisis and its aftermath, the Fed used a vari-ety of new facilities to provide liquidity to the financial system aswell as unconventional tools, such as large-scale asset purchases. Asa result, reserves in the banking system have reached unprecedentedlevels. Marking a significant shift in its policy framework, the Fedstarted paying interest on reserves (IOR) in October 2008 to controlrates in an environment of superabundant reserves.

Elevated reserves and the new monetary policy tools affectedtrading dynamics of money market participants. In the federal fundsmarket, government-sponsored enterprises (GSEs), which are noteligible to earn IOR, became the primary lenders, while large andforeign banks started borrowing funds at rates below IOR for arbi-trage purposes. Mainly because of this fragmented market structure,the IOR could not set a lower bound on the federal funds rate, lead-ing the Fed to introduce a supplementary tool, the ON RRP facility,in September 2013 to enhance monetary control.

The changing regulatory environment also created new incen-tives for market participants amid a substantial decline in the lever-age of securities dealers. Among the new regulations, the change inthe assessment base for the Federal Deposit Insurance Corporation(FDIC) deposit insurance and the Basel III leverage ratio require-ments are of particular importance. The former made wholesalefunding more costly for U.S.-chartered banks relative to that of U.S.


branches and agencies of foreign banks, incentivizing domestic banksto reduce their money market borrowing. The latter incentivized for-eign banks to dynamically deleverage through money market activ-ity given regional differences in implementation of the leverage ratio.Meanwhile, both leverage levels and net repo liabilities of the broker-dealer sector decreased notably, creating an important contrast tothe pre-crisis period, during which such institutions largely operatedoutside of the regulated banking system.

The remainder of this section describes the pre- and post-crisismonetary policy and regulatory frameworks in more detail to giveperspective on the empirical analysis that follows. We cover differentsegments of the money market in sections 2.1 and 2.2, provide back-ground on the Fed’s monetary policy implementation framework insection 2.3, and discuss regulations in section 2.4.

2.1 Bank Reserves and Activity in the Federal Funds Market

Banks are required to maintain a minimum level of reserves at theFederal Reserve Banks in their Districts.2 Historically, banks avoidedholding excess reserves, as such balances did not earn any interest.Indeed, total reserve balances in the banking system averaged about$10 billion in 2007, with only about $2 billion in excess reserve bal-ances, while total bank assets were close to $10 trillion over the sameperiod. As can be seen in figure 1, reserves in the system increasedto more than $800 billion at the end of 2008, as the Fed providedliquidity during the financial crisis through several facilities.3 Follow-ing subsequent rounds of asset purchases from 2009 to 2014, totalreserve balances averaged nearly $2.5 trillion in 2016.4

2We will be referring to depository institutions with reserve accounts simplyas banks. See Regulation D Reserve Requirements for a full list of financial insti-tutions in this category, available at https://www.federalreserve.gov/boarddocs/supmanual/cch/int depos.pdf.

3See https://www.federalreserve.gov/monetarypolicy/bst crisisresponse.htmfor details on the Fed’s crisis response, and https://www.federalreserve.gov/monetarypolicy/expiredtools.htm for a list of expired liquidity provision facilities.

4Between November 2008 and October 2014, the Fed purchased nearly $1.7trillion in Treasury securities and about $2 trillion in agency mortgage-backedsecurities, as well as $170 billion in agency debt securities in order to put down-ward pressure on longer-term interest rates. See Krishnamurthy and Vissing-Jorgensen (2011) and D’Amico et al. (2012) for a discussion of the economicrationale and effects of large-scale asset purchases.


Figure 1. Reserves and Federal Funds

Note: Data are quarterly and obtained from Call Reports.

The unprecedented increase in reserve balances changed the land-scape for the federal funds market. Federal funds are unsecured loansof reserve balances between banks and other eligible institutions,mainly GSEs. Federal funds transactions are typically conductedfor an overnight term and are carried out either directly betweenthe institutions or through third-party brokers. Historically, trans-actions in the federal funds market facilitated the redistribution ofreserve balances, whereby banks with reserve balances in excess ofthe required levels lent to banks in need of reserves. The surge inreserve balances led to a substantial decline in banks’ need for short-term borrowing to cover idiosyncratic funding shortfalls. To ensuremonetary control and promote efficiency in the banking system, theFed introduced the IOR as a new monetary policy tool in October2008. As a result, incentives for banks to lend federal funds at ratesbelow the IOR rate were largely eliminated.

As shown in figure 1, the end-of-quarter outstanding amount offederal funds borrowed by banks declined to roughly one-fourth ofthe level observed prior to the global financial crisis by 2011, andit has remained low since then. Moreover, volume in the overnightfederal funds market declined from an average of $200 billion perday in 2007 to $60 billion per day at the end of 2012, according to


estimates provided by Afonso, Entz, and LeSueur (2013b). Banksthat used to lend more than half of the federal funds before the cri-sis accounted for only a fraction of the lending activity after 2008.GSEs, specifically the Federal Home Loan Banks, which are not eli-gible to earn IOR, have been the main lenders in the post-crisisperiod.5 On the borrowing side, Afonso, Entz, and LeSueur (2013a)show that mostly banks under the umbrella of bank holding compa-nies and foreign banking organizations have been purchasing federalfunds from GSEs for arbitrage purposes.6 These institutions bor-row federal funds at rates below the IOR and place the cash intheir reserve accounts to earn the spread between the IOR and thefederal funds rate. These transactions have been relatively more prof-itable for foreign banks, as they are not subject to assessment by theFDIC.7

The changing landscape in the federal funds market raises theimportant question of whether pass-through from the federal fundsrate to other overnight rates has been affected over time. In addi-tion, superabundant reserves may have implications for cash flowsin the market associated with days of reserve maintenance period.8

In the pre-crisis era, activity in the federal funds market was inpart driven by maintenance-period dynamics, as shown by Spindtand Hoffmeister (1988), Griffiths and Winters (1995), and Hamilton(1996), among others. However, superabundant reserves, combinedwith the finding by Ennis and Wolman (2015) that reserves in thesystem have been fairly widely distributed across banks since mid-2009, suggest that calendar effects associated with reserve mainte-nance have likely dissipated in the post-crisis era.

5See Ashcraft, Bech, and Frame (2010) for a detailed description of the FederalHome Loan Bank System and its role as a liquidity provider to banks.

6In the current context, foreign banking organizations are U.S. branches andagencies of foreign banks.

7In 2011, the FDIC changed the assessment base for its deposit insurancescheme from domestic deposits to total assets minus equity, making larger bal-ances more costly for domestic banks regardless of funding source. See Kreicher,McCauley, and McGuire (2013) for a detailed discussion.

8See the Reserve Maintenance Manual for reporting requirements as wellas calculation and maintenance of reserve balances, available at http://www.federalreserve.gov/monetarypolicy/files/reserve-maintenance-manual.pdf.


2.2 Other Money Market Segments

The Fed’s monetary policy implementation framework relies ontargeting the federal funds rate and cross-market linkages acrossmoney markets. Specifically, overlapping participants in variousmoney market segments and arbitrage activity ensure strong co-movement that the Fed depends on for effective monetary policytransmission.

The Eurodollar market is another segment for unsecured fundingthat is broader than the federal funds market. Eurodollars are U.S.dollar-denominated deposits held in a bank or a bank branch locatedoutside of the United States. U.S. banks and foreign banking orga-nizations cannot directly borrow in the Eurodollar market but cantake Eurodollar deposits, mainly through their Caribbean branches,and transfer them onshore to fund U.S. operations. Eurodollardeposits that remain outside the United States are not coveredby FDIC deposit insurance, while those that are transferred to aninsured U.S. affiliate are included in the deposit insurance assess-ment base. Because of their unsecured nature and regulatory treat-ment, Eurodollar deposits constitute a close substitute to federalfunds. However, the Eurodollar market has a more diverse set ofparticipants than the federal funds market, as participants do nothave to have an account at the Fed. Cipriani and Gouny (2015)estimate that the average volume in the brokered Eurodollar mar-ket is three to four times larger than that in the brokered federalfunds market. However, there has been a substantial drop in theEurodollar volume following the money fund reform compliance datein October 2016, as prime funds pulled back from lending in thismarket.

The major segment of the money market for secured funding isthe repo market. A repo is effectively a collateralized loan in whichthe lender of the cash receives the security as collateral, and the bor-rower pays the lender interest on the loan. Sale of securities takesplace under an agreement to repurchase them at a specified priceon a later date.9 The repo market can broadly be divided into two

9Fed transactions in the repo market are defined from the point of view ofthe market participants, that is, a collateralized transaction in which the Fedborrows cash is called a reverse repo.


parts: the bilateral market where the two parties interact directly,and the triparty market where clearing/brokerage services of a thirdparty are involved. Total volume of the Treasury repo market iswell above $2 trillion. Copeland et al. (2014) and Baklanova et al.(2016) provide estimates of volumes in different repo market seg-ments. Cash borrowers (or securities lenders) in the repo marketinclude banks and securities dealers, while money market mutualfunds (or money funds) and GSEs are among the primary lenders ofcash (or borrowers of securities).

The final segment of the money market we consider is the com-mercial paper market, in which large corporations issue debt for afixed maturity. Many companies issue commercial paper when theyneed to raise short-term cash, as it is a lower-cost alternative to bankloans. Although commercial paper is unsecured, it is considered avery safe investment, as typically only creditworthy companies withhigh credit ratings issue such securities. Commercial paper is espe-cially attractive for institutional investors, as they are liquid andhave a low risk of default.

2.3 Monetary Policy Implementation Framework

Historically, adjustment of the level of reserve balances in the bank-ing system to move the effective federal funds rate toward the tar-get level set by the Federal Open Market Committee (FOMC) wasthe central pillar of monetary policy implementation. Given scarcereserve balances in the system, the Fed would affect the federal fundsmarket rate by announcing a target level and managing the amountof reserves available to the banking system through open marketoperations (OMOs). These operations would influence the rate inthe federal funds market, where banks experiencing shortfalls couldborrow from banks with excess reserves. Given the low volume ofreserves at the Fed, around $10 billion, even small OMOs couldsignificantly affect the market rate. Changes in the federal fundsrate would then be transmitted to other short-term interest rates,to longer-term interest rates, and eventually to inflation and eco-nomic activity. This framework worked seamlessly while the Fedwas operating with a balance sheet of less than $1 trillion before thecrisis.


The global financial crisis precipitated changes in the operationalframework of the Fed.10 In an environment with superabundantreserves, the conventional approach based on changing the quantityof reserves via OMOs would not work. As a result, the Fed extendedits monetary policy toolkit. In the fall of 2008, the Fed started payinginterest on banks’ reserve balances, which became the primary toolof its new monetary policy implementation framework in controllingshort-term interest rates.

Although adjusting the IOR rate is an effective way to movemarket interest rates in an environment of superabundant reserves,federal funds have generally traded below this rate, mainly due tothe fact that only banks can earn IOR. GSEs, the other major groupof participants in the federal funds market, still have an incentive tolend at rates below the IOR rate, as they do not receive interest ontheir reserve accounts. Moreover, FDIC fees and other balance sheetconstraints, such as capital and liquidity regulations, limit arbitrageactivity by banks that would push the market rate toward the IORrate.

In order to enhance monetary control and put an effective floorunder short-term interest rates, the Fed introduced the ON RRPfacility as a supplementary tool for its implementation of monetarypolicy. ON RRPs are offered to a broader set of financial institu-tions, including money funds that do not have access to the federalfunds market. In September 2014, the FOMC issued a statementsummarizing the new operating strategy, and in December 2015, itsuccessfully lifted the federal funds rate from its near-zero range inthis framework.11

The primary tool of the new operating framework, IOR, hasimportant implications for the transmission of monetary policy fromfederal funds to the repo market. In the pre-crisis era, the active pres-ence of large banks in both the federal funds and repo markets wascrucial to the co-movement of these two rates. The unsecured natureof the federal funds transactions has typically resulted in a small and

10See Ihrig, Weinbach, and Meade (2015) for an in-depth discussion of theevolution of the Fed’s monetary policy implementation framework through thefinancial crisis and its aftermath.

11See https://www.federalreserve.gov/monetarypolicy/policy-normalization.htm for further details on policy normalization. Anderson et al. (2016) providean overview of money market developments after the liftoff.


positive spread between the federal funds rate and the rate on repotransactions where the underlying collateral is a U.S. Treasury oragency security. However, market-determined, or effective, federalfunds rate staying below the repo rates became a frequent phenom-enon amid superabundant reserves and the ELB on the funds rate.The negative spread reflects reduced scope for arbitrage activity dueto IOR, aside from the dramatic reduction in banks’ needs for short-term borrowing, as discussed previously. Specifically, when the reporates were greater than the federal funds rate in the past, banks couldborrow in the federal funds market and place the cash in the repomarket, creating downward pressure on the repo rates and pushingthe effective federal funds rate up. However, in the presence of theIOR, the incentive for banks to engage in arbitrage activity acrossthe federal funds and repo markets exists only when the repo ratesare above the IOR. Although GSEs may also engage in this type ofarbitrage, frictions—such as internal restrictions or intraday timingconsiderations—likely limit such activity. As a result, we expect aweaker link between the effective federal funds rate and the reporates in the ELB sample, on net.

The supplementary monetary policy tool of the new framework,the ON RRP facility, has also been affecting overnight fundingdynamics since its inception in September 2013. The Fed has beenconducting operations on a daily basis at a pre-announced offer-ing rate. Through this facility, the Fed borrows cash from eligiblecounterparties in exchange for Treasury securities from its open mar-ket portfolio. These operations provide an investment vehicle formoney market participants who usually compare the facility’s offer-ing rate with rates in the market and determine whether to bid inthe operation.

The ON RRP operations are, in essence, similar to the tempo-rary OMOs in the form of reverse repos conducted by the Fed priorto the crisis, but there are also important differences. Participa-tion in the ON RRP operations is open to a wide range of entities,including money funds, banks, and GSEs, in addition to primarydealers of the Fed. Frost et al. (2015) show that money funds havebeen the dominant cash lenders in ON RRP operations. By expand-ing the set of alternative investments available to money funds andGSEs, the ON RRP was intended to strengthen rate control. Thesecond important difference between the ON RRP and conventional


Figure 2. ON RRP Operations

Note: Data are daily and obtained from the Federal Reserve Bank of New York,available at https://apps.newyorkfed.org/markets/autorates/temp.

temporary OMOs is that the latter was conducted to move the effec-tive federal funds rate close to the FOMC’s target, while the formeris intended to set a floor for the overnight rates. The mechanism issimilar to that of IOR for banks in the federal funds market; ONRRP counterparties do not have an incentive to invest in alternativesources unless they are offered the ON RRP rate or higher. Indeed,Potter (2015) shows that the ON RRP has established a soft floor,as the FOMC intended—that is, although some trades likely occurbelow the ON RRP rate, volume-weighted average overnight fundingrates have mostly been above the offering rate. A general reductionin the volatility of overnight rates is another expected effect of thesoft floor set by the ON RRP. Such effects are especially impor-tant on financial reporting days when borrowers contract the size oftheir balance sheets, leaving cash lenders looking for alternative safeinvestment options.

Take-up at the ON RRP facility trended up for about a year fol-lowing its inception in September 2013, as can be seen from figure 2.One year later, the FOMC reduced the overall limit on the facilitysubstantially (from $1.4 trillion to $300 billion) and introduced anauction process to allocate reverse repos in the event that the overalllimit is binding. This change led to a sharp drop in money market


rates on that quarter-end, as it left cash lenders scrambling for alter-native investments. In October 2014, term RRP operations spanningyear-end were announced. These operations were conducted overquarter-ends until December 2015 and helped address downwardpressure on rates, suggesting that perceived investment capacity isan important factor in determining the effectiveness of reverse reposin supporting rates.

At the time of the rate hike in December 2015, the aggregatecap on ON RRP operations was temporarily suspended. Currently,ON RRP operations are limited only by the value of Treasury secu-rities in the Fed’s open market portfolio that are available for theseoperations, which stand around $2 trillion.

2.4 New Banking Regulations and Dealer Leverage

The announcement and implementation of Basel III capital andliquidity reforms had a significant effect on the post-crisis financiallandscape. In terms of their effects on money markets, the liquiditycoverage ratio (LCR) and the leverage ratio are of particular interestamong the Basel III reforms.

The LCR rule, which was first proposed in the United Statesin October 2013, requires banks to hold high-quality liquid assets(HQLAs) to meet cash outflows under a thirty-day stress scenario.Therefore, it has potential implications for bank activity in overnightmoney markets, as many assets and liabilities closely tied to thesemarkets are under the jurisdiction of the LCR.

In the LCR calculation, cash inflows and outflows over a thirty-day stress period are aggregated and netted with specific rates fordifferent assets and liabilities. In some cases, lending in the federalfunds market may decrease the LCR of the lending bank. Althoughlending federal funds reduces the LCR numerator because reservesare counted as HQLAs, cash inflow assumptions for financial insti-tutions typically imply limited impact of such activity on LCR. Sim-ilarly, lending in the repo market in which the underlying collateralis in the HQLA category has no effect on a bank’s LCR. On theborrowing side, funding non-HQLAs through unsecured interbankborrowing or repos deteriorates LCR, and thus incentivizes banks toreduce their reliance on such financing. By the time the initial LCRannouncement was made, banks had already reduced their reliance


on wholesale funding substantially (see, for example, Choi and Choi2016). Finally, IOR arbitrage trades increase a bank’s LCR, as theborrowed cash is parked in the arbitrageur bank’s reserve account—which is treated as HQLAs with no haircuts—and the cash out-flow assumption associated with borrowing results in a less-than-proportional increase in the denominator. All told, we do not expectthe effect of the LCR to be material for overnight money marketdynamics.

Another notable aspect of Basel III for money market activityhas been the leverage ratio requirement, which requires banks tohold tier 1 equity equivalent to at least 3 percent of their lever-age exposure calculated using their on- and off-balance-sheet assets,including reserves. The Supplementary Leverage Ratio, the regu-lation that implements the leverage ratio provisions in the UnitedStates, bases the relevant calculations on averages of daily valuesfor on-balance-sheet items. In contrast, for most foreign banks, dis-closures are based on month- or quarter-end levels. This regionalimplementation difference incentivized foreign banks to contracttheir balance sheets on financial reporting days and expand onother days. The leverage ratio requirements were announced inmid-2013, and banks started disclosing their leverage ratios to thepublic in January 2015, including three quarters of historical data.Becoming compliant before the beginning of public disclosures wasan important motivation for banks to adjust their balance sheets.Therefore, we expect stronger reporting-day effects on rates inthe ELB sample following the introduction of the leverage ratiorequirements.

Declining leverage of securities broker-dealers has been anotherimportant feature of the post-crisis landscape (Adrian et al. 2013).Dealers were not subject to leverage limits prior to the crisis, as theywere outside the regulated banking system. However, four out of thefive major standalone investment banks with dealer arms have beenintegrated into bank holding companies via either acquisitions orconversions. This change has been among the main drivers of lowerdealer leverage along with generally increased risk aversion in theaftermath of the crisis.

Dealers dynamically adjust their balance sheets mainly throughshort-term borrowing in the form of repos, as discussed in Adrianand Shin (2010). Along with overall leverage, repo activity of


Figure 3. Repo Financing Activity by Securities Brokersand Dealers

Notes: Data are annual and obtained from the Financial Accounts of theU.S. statistical release (Z.1) of the Federal Reserve Board, available athttp://www.federalreserve.gov/releases/z1/. TA (TL) denotes total assets (lia-bilities), RA (RL) denotes repo assets (liabilities), and NR is net repo financing,that is, RL − RA.

dealers also declined relative to the pre-crisis norms. As can be seenfrom figure 3, although repo lending by dealers has been relativelystable since 2001, their repo borrowing has been lower since 2007.The change in net repo financing has been substantial: the ratio ofnet repo liabilities to total liabilities for dealers has been steadilydecreasing since its peak in 2007, and reached about 8 percent in2015, almost one-fourth of its level in 2007. Meanwhile, the ONRRP facility limited downward pressure on the repo rates aroundfinancial reporting days by setting a soft floor. Therefore, we expectweaker quarter-end effects on repo rates in the ELB sample. Table 1summarizes all the aforementioned changes in the monetary policyand the regulatory environment, as well as their anticipated effectson overnight money markets.

3. Data

Our sample covers two main periods: the pre-crisis period that spansfrom January 2, 2001, to July 31, 2007; and the ELB period that runs


Table 1. Changes in Monetary and RegulatoryPolicy and Implications

Changes in Monetary Anticipated Effects onand Regulatory Policy Overnight Money Markets

Superabundant Reserves and IOR

Lower trading volumes in thefederal funds market

(i) Weaker co-movement of EFFRwith other rates

(ii) Increased EFFR volatility

Reduced scope for EFFR-RPRarbitrage trades by banks

Weaker EFFR-RPR co-movement

Widespread distribution ofreserves

MP effects diminish in theaggregate

ON RRP

Inclusion of money funds andGSEs among counterparties

(i) Stronger co-movement ofovernight interest rates

(ii) Lower interest rate volatility(iii) Weaker financial reporting

effects on RPR

New Regulations and Lower Dealer Leverage

LCR IOR arbitrage trades moreattractive, but limited effect dueto other regulatory constraints

FDIC assessment base changeLeverage ratio

Stronger financial-reporting-dayeffects on unsecured rates andtheir volatility

Diminishing leverage and repofinancing by dealers

Weaker financial-reporting-dayeffects on RPR

from December 17, 2008, to August 28, 2015.12 The former is asso-ciated with the conventional monetary policy operating frameworkand serves as a benchmark, while the latter is a period during which

12We exclude the period from mid-2007 to late 2008 from our analysis, as thisperiod is associated with unprecedented movements in the rates driven by thefinancial crisis.


overnight money markets were subject to the significant changesdiscussed above.

Our data set consists of four overnight money market interestrates. The first one is the effective federal funds rate (EFFR), whichis calculated as a volume-weighted average of rates on brokered fed-eral funds trades and published by the Federal Reserve Bank ofNew York (FRBNY). The second one is the Eurodollar rate (EDR),which represents the cost of alternative unsecured funding for largebanks. We use the EDR data that the FRBNY started collectingin March 2010. Prior to this date, we use the data obtained fromWrightson ICAP in the ELB sample. For our pre-crisis analysis, wesubstitute the EDR with the overnight London interbank offeredrate (LIBOR), obtained from Bloomberg, because Eurodollar dataare not available for that period. LIBOR is a commonly used indi-cator for the average rate at which banks may get short-term loansin the London interbank market, and it serves as reference rate forvarious debt instruments.13 The third key rate in our analysis is arepresentative rate of secured funding from the repo market. We usethe volume-weighted average rate for Treasury GC (general collat-eral) repo obtained from the FRBNY, which we will refer to as RPR.Finally, we use the overnight AA non-financial commercial paperrate (CPR) released by the Federal Reserve Board.14 An importantfeature of the CPR in our context is that it represents an unsecuredfunding rate that is not directly affected by the changing monetarypolicy framework and new banking regulations discussed above.

Visual investigation suggests very strong co-movement amongthe rates during normal times (figure 4). Moreover, the sample meansand standard deviations of the rates are remarkably close in thepre-crisis period, as shown in panel A of table 2. However, as onecan infer from figure 5 and panel B of table 2, the co-movement ofrates appears to have weakened over the ELB period, on net. Forexample, RPR remained especially elevated relative to unsecuredrates around late 2011, reportedly due to longer dealer position-ing in Treasury securities that coincided with the Fed’s Maturity

13See Hou and Skeie (2014) for a detailed description of the rate-setting mech-anism and efforts to reform the LIBOR.

14Data are available at http://www.federalreserve.gov/releases/cp/.


Figure 4. Overnight Money Market Interest Rates:2001–07

Notes: Data are daily. EFFR and RPR are from FRBNY. LIBOR is fromBloomberg. CPR is from the commercial paper data release of the Federal ReserveBoard.

Extension Program as well as higher Treasury debt issuance.15 Inaddition to weaker co-movement, calendar effects relative to the levelof the rates seem stronger, on average, over the ELB period, and thesample moments also show more variation across the rates. In thenext section, we lay out the empirical framework to quantify suchdifferences and analyze them in detail.

Another difference between the two samples is the degree of sta-tionarity of the rates. In table 3, we report the augmented Dickey andFuller (1979) (ADF) unit-root test statistic and the Elliott, Rothen-berg, and Stock (1996) (ERS) point-optimal test statistic, which hashigher power against persistent alternatives. In the pre-crisis sample,we cannot reject the null of a unit root in the interest rates at anyconventional significance level with respect to both test statistics. In

15During this program, the Fed sold about $650 billion of short-term securitiesand used the proceeds to buy longer-term securities. By extending the averagematurity of its securities portfolio, the Fed aimed to put downward pressure onlonger-term interest rates to ease conditions in financial markets.


Table 2. Descriptive Statistics of Money Market Rates

LIBOR/EFFR RPR EDR* CPR

A. January 2, 2001–July 31, 2007

Mean 2.937 2.881 2.999 2.927St. Dev. 1.660 1.639 1.661 1.66210th 1.010 0.980 1.058 0.99050th 2.480 2.440 2.541 2.45090th 5.250 5.220 5.301 5.250AC(1) 0.999 0.999 0.999 0.999

B. December 17, 2008–August 28, 2015

Mean 0.129 0.118 0.137 0.107St. Dev. 0.042 0.068 0.051 0.05810th 0.080 0.030 0.080 0.04050th 0.130 0.110 0.130 0.09090th 0.190 0.210 0.210 0.190AC(1) 0.954 0.920 0.950 0.958

Notes: Data are daily. Mean standard deviation, and quantiles are reported in per-cent. AC(1) denotes first-order autocorrelation.*LIBOR is used for panel A calculations and EDR is used in panel B.

Figure 5. Overnight Money Market Interest Rates:2008–15

Notes: Data are daily. EFFR and RPR are from FRBNY. EDR is from FRBNYafter March 2010, and from Wrightson ICAP prior to this date. CPR is from thecommercial paper data release of the Federal Reserve Board.


Table 3. Unit-Root Tests

LIBOR/EFFR RPR EDR* CPR

A. ADF Test

Pre-crisis −1.24 −1.31 −1.00 −1.02ELB −3.37 −3.17 −2.82 −2.52

B. ERS Test

Pre-crisis 251.3 275.6 158.2 195.7ELB 1.2 2.8 2.5 3.4

Notes: ADF is the augmented Dickey and Fuller (1979) test with the 1, 5, and 10percent critical values of −3.44, −2.87, and −2.57, respectively. ERS is the point-optimal test of Elliott, Rothenberg, and Stock (1996) with the 1, 5, and 10 percentcritical values of 1.99, 3.26, and 4.48, respectively.*LIBOR is used for panel A calculations and EDR is used in panel B.

contrast, we reject the null of unit root for all rates in the ELB sam-ple according to the ADF test statistics, with the exception of CPR,and for all rates according to the ERS test statistic. Therefore, theinterest rates are well approximated by integrated processes with alikely common stochastic trend in the pre-crisis sample, reflectingthe fact that this period contains a full monetary policy cycle witheasing early in the period followed by a gradual tightening begin-ning in 2004. In the ELB period, the rates are persistent but notintegrated against the backdrop of no change in the target federalfunds rate. Our modeling strategy incorporates this difference in ratedynamics.

4. Models

We specify models that account for co-movement and persistence ofthe money market rates as well as time variation in their volatili-ties and cross-correlations. We also allow for various calendar factorsthat likely affect dynamics of rates on specific days. The unit-roottests suggest that the interest rates are well approximated by inte-grated processes in the pre-crisis sample while they are persistent but


stationary during the ELB sample. Therefore, we estimate differentmodels for the pre-crisis and the ELB period.

The pre-crisis model is a vector error correction (VEC) processthat incorporates the long-run equilibrium relationship of overnightmoney market rates. Let yt denote the vector of the interest ratesat time t, that is, yt = (EFFRt,RPRt,LIBORt,CPRt)′ in the pre-crisis sample. The interest rate dynamics are characterized by thefollowing VEC model:

Δyt = Adt + βΔTFFRt +p∑

j=1

ΦjΔyt−j + Θzt−1 + εt, (1)

where dt is a vector of indicator variables for calendar effects, whichwe will explain in detail; TFFR is the target federal funds rate; zt

is a vector of error correction terms; and εt is a zero-mean martin-gale difference vector process, which is possibly heteroskedastic.16

Reflecting the pre-crisis monetary policy operating framework, weimpose the restriction that there are three distinct co-integratingrelationships and that all of them involve EFFR. Formally, we havezit = y1t − (ci + γiyi+1,t), where i = 1, 2, 3.17

The vector of calendar effects, dt, contains ten indicator vari-ables to account for reserve maintenance period days: two indicatorsfor elevated payment days within a month (15th and 25th), two forfinancial reporting days (month-end and quarter-end), and a dummyvariable to control for the brief disturbance in money markets causedby the September 2001 terror attacks. As a result, the model doesnot contain a constant vector because it cannot be separately iden-tified given the set of maintenance period indicators. We set p = 4based on the Schwarz information criterion.18

There exists a mapping from this VEC system to a VAR thatcan be defined for the level of interest rates. This mapping allows

16Alternatively, one can also estimate a VAR for the level of the rates assumingstationarity, with TFFR on the right-hand side. We find that this alternative spec-ification yields qualitatively the same results. The full set of results is availablefrom the authors upon request.

17We obtain nearly identical results when we estimate the number of co-integrating relationships as well as the co-integration parameters in a lessrestricted fashion as in Johansen (1995).

18The total number of parameters to be estimated is equal to 140, which resultsin approximately forty-six observations per parameter.


us to directly compare the results from the pre-crisis period withthose from the ELB period, as the model for the latter sample is aVAR in levels. Let Ψj for j = 1, . . . , p + 1 denote the autoregres-sive coefficient matrices in the implied VAR in levels. Then we haveΨ1 = Φ1+I+ΘΓ, where I is an identity matrix, Γ = (i,−diag{−γ}),i is a vector of ones, γ is the vector of co-integration slopes givenpreviously, and diag(.) indicates a diagonal matrix, Ψj = Φj −Φj−1for j = 2, . . . , p, and Ψp+1 = −Φp.19

For the ELB period, we specify the following VAR model in levelsgiven the stationary behavior of interest rates in this sample:

yt = Πdt +p∑

j=1

Ξjyt−j + εt, (2)

where dt is now a 9 × 1 vector that contains month-end, quarter-end, day-of-week, and elevated payment flow-day indicators.20 Notethat the EDR replaces the LIBOR in this sample, so that yt =(EFFRt, RPRt, EDRt, CPRt)′. We set p = 3 based on Schwarzmodel-selection criteria.21

Both visual investigation and formal testing of the model resid-uals suggest significant volatility clustering in both sample periods.Hence, we model volatility dynamics using multivariate GARCHmodels. Our modeling strategy closely follows that of Bollerslev(1990); however, instead of assuming a constant conditional correla-tion matrix, we allow for different correlation structures on financialreporting days. Therefore, our specification can be thought of as ahybrid of the constant correlation model and the dynamic corre-lation model of Engle (2002), who postulates a fully time-varyingconditional correlation matrix. Let E(εtε

′t|Ωt−1) = Ht, where Ωt is

the information set at time t; then we can write

Ht = DtRtDt, (3)

19A caveat is that in the pre-crisis model, shocks are permanent due to themodeling of interest rates as integrated processes.

20Day-of-week indicators replace those for maintenance period days, as thelatter become insignificant amid abundant reserves in the ELB period.

21This model has eighty-four parameters to be estimated, resulting in seventy-eight observations per parameter.


where Dt = diag{√

hit

}, hit = V ar(εit|Ωt−1), and Rt =

Corr(εt|Ωt−1). The individual variances are modeled via the fol-lowing GARCH specification:

hit = ωi + τiε2i,t−1 + δihi,t−1 + λi,1Im,t + λi,2Iq,t, : i = 1, . . . , 4, (4)

where Im and Iq are month-end and quarter-end indicators, respec-tively. In this specification, the variance at time t is essentially aweighted average of its lagged value, the new information at time t−1that is captured by the most recent squared residual, the long-rununconditional variance, and the level shifts in volatility on financialreporting dates. We estimate the GARCH equation under variancetargeting so that ωi is a function of the sample variance of εi,t and themean vector of the indicator series. Finally, the correlation matrixRt is specified as follows:

Rt = Im,tRm + Iq,tRq + (1 − Im,t − Iq,t)Rn, (5)

where Rm, Rq, and Rn are correlation matrices of GARCH resid-uals, that is, h

−1/2it εt, at month-ends, quarter-ends, and all other

days, respectively.

5. Empirical Results

5.1 Co-movement of Rates and Monetary PolicyTransmission

Our estimates for the pre-crisis sample are consistent with the con-ventional monetary policy implementation framework. As shown inpanel A of table 4, lagged EFFR variables are significant in allother rate equations, implying that interest rates were adjusting inresponse to changes in the policy rate. Moreover, the EFFR was notresponding to the other rates, as indicated by the insignificance oflagged rates in the first column. As can be seen in panel B, changesin the target federal funds rate are highly significant in all equa-tions of the VEC model.22 Other than the EFFR, no other interest

22To save space, we do not report the parameter estimates of ci and γi in theco-integration equation. These estimates are highly statistically significant andvery close to 0 and −1 in all equations, respectively.


Table 4. Overnight Money Market Rates beforethe Financial Crisis

EFFR RPR LIBOR CPR

A. Autoregressive Terms (Sum)

EFFR 0.947 0.449 0.345 0.521(0.00) (0.00) (0.00) (0.00)

RPR 0.033 0.694 0.010 0.001(0.39) (0.00) (0.72) (0.98)

LIBOR 0.016 −0.125 0.546 −0.091(0.91) (0.41) (0.00) (0.43)

CPR 0.005 −0.021 0.099 0.570(0.97) (0.92) (0.21) (0.00)

B. Other Variables

ΔTFFR 0.454 0.406 0.337 0.416(0.00) (0.00) (0.00) (0.00)

15th 5.50 6.04 6.10 7.00(0.00) (0.00) (0.00) (0.00)

25th 4.33 0.69 0.09 1.14(0.00) (0.24) (0.75) (0.00)

Month-End 5.33 4.15 7.43 6.20(0.00) (0.00) (0.00) (0.00)

Quarter-End 5.66 −12.52 17.33 11.17(0.03) (0.01) (0.00) (0.00)

Notes: Columns represent equations of the models. The sum of autoregressive termscorrespond to ΣΨj in terms of the notation of section 4. p-values based on robust(HAC) standard errors are reported in parentheses. Calendar effects are in basispoints. Daily sample runs from January 2, 2001, to July 31, 2007.

rate in the system had predictive power for the remaining inter-est rates. Overall, these results confirm that overnight funding rateswere adjusting in response to policy intervention and dynamics inthe federal funds market. These results are consistent with the viewthat the overnight money markets were tightly connected throughthe federal funds market in the pre-crisis period.

The estimates from the ELB sample shown in panel A of table 5paint a different picture. The federal funds and Eurodollar marketsappear to be closely connected, as indicated by the statistical and


Table 5. Overnight Money Market Rates at the ELB

EFFR RPR EDR CPR


EFFR 0.911 0.107 0.223 0.153(0.00) (0.21) (0.00) (0.02)

RPR 0.032 0.809 0.14 −0.011(0.00) (0.00) (0.29) (0.47)

EDR −0.024 0.048 0.705 0.000(0.53) (0.50) (0.00) (1.00)

CPR 0.036 0.054 0.069 0.881(0.01) (0.14) (0.00) (0.00)

B. Other Variables

15th 0.80 3.29 0.85 0.96(0.00) (0.00) (0.00) (0.00)

25th −0.26 0.65 −0.08 0.37(0.01) (0.08) (0.36) (0.05)

Month-End −0.14 3.47 −0.13 0.37(0.63) (0.00) (0.72) (0.20)

Quarter-End −3.21 −0.41 −5.07 −1.58(0.00) (0.70) (0.00) (0.03)

Notes: Columns represent equations of the models. The sum of autoregressive termscorrespond to ΣΞj in terms of the notation of section 4. p-values based on robust(HAC) standard errors are reported in parentheses. Calendar effects are in basispoints. Daily sample runs from December 17, 2008, to August 28, 2015.

economic significance of the EFFR coefficients in the EDR equa-tion. Similarly, the EFFR is linked to the CPR, which is another keyunsecured rate in the system, although to a lesser extent than theEDR. These results imply that the EFFR continued to be an anchorfor unsecured rates in the ELB period, although its transmissionweakened relative to pre-crisis norms, especially in the case of theCPR.

The most dramatic change across the two periods concerns thetransmission from the federal funds to the repo market. The EFFRno longer predicts RPR movements in the ELB period. This dis-connect between the two rates that used to move in tandem also


Table 6. Overnight Money Market Rates beforethe ON RRP

EFFR RPR EDR CPR


EFFR 0.911 0.112 0.226 0.164(0.00) (0.26) (0.00) (0.03)

RPR 0.018 0.803 0.002 −0.022(0.13) (0.00) (0.88) (0.23)

EDR −0.002 0.045 0.739 0.005(0.97) (0.59) (0.00) (0.94)

CPR 0.028 0.052 0.047 0.880(0.04) (0.20) (0.01) (0.00)

B. Calendar Effects

15th 1.11 3.93 1.26 1.42(0.00) (0.00) (0.00) (0.00)

25th −0.29 0.70 0.01 0.54(0.03) (0.17) (0.94) (0.02)

Month-End 0.79 3.94 1.19 0.71(0.00) (0.00) (0.00) (0.06)

Quarter-End −3.14 −1.06 −4.29 −2.02(0.00) (0.45) (0.00) (0.04)

Notes: Columns represent equations of the models. The sum of autoregressive termscorrespond to ΣΞj in the notation of section 4. p-values based on robust (HAC) stan-dard errors are reported in parentheses. Calendar effects are in basis points. Dailysample runs from December 17, 2008, to September 20, 2013.

emphasizes the diminished role of banks as arbitrageurs, as discussedin section 2.1. Overall, we conclude that co-movement of the EFFRwith other rates became noticeably weaker in the ELB sample amidsuperabundant reserves, subdued trading, and dominance of IORarbitrage trades in the federal funds market.

To assess the effects of the ON RRP facility on money marketdynamics, we now focus solely on the ELB period and estimate VARmodels for the two subsamples separated by the inception of theON RRP facility in September 2013. The comparison of the resultssummarized in tables 6 and 7 suggests that the ON RRP had two


Table 7. Overnight Money Market Rates afterthe ON RRP

EFFR RPR EDR CPR


EFFR 0.823 0.033 0.290 0.333(0.00) (0.83) (0.06) (0.00)

RPR 0.102 0.813 0.096 0.084(0.00) (0.00) (0.00) (0.00)

EDR −0.127 0.101 0.416 −0.145(0.01) (0.42) (0.00) (0.01)

CPR 0.129 0.126 0.113 0.565(0.04) (0.21) (0.12) (0.00)

B. Calendar Effects

15th −0.02 1.61 −0.18 −0.24(0.87) (0.00) (0.28) (0.27)

25th −0.08 0.46 −0.22 −0.12(0.50) (0.17) (0.09) (0.54)

Month-End −2.35 2.42 −3.25 −0.37(0.00) (0.00) (0.00) (0.15)

Quarter-End −3.41 1.10 −6.80 −1.05(0.00) (0.37) (0.00) (0.02)

Notes: Columns represent equations of the models. The sum of autoregressive termscorrespond to ΣΞj in the notation of section 4. p-values based on robust (HAC) stan-dard errors are reported in parentheses. Calendar effects are in basis points. Dailysample runs from September 23, 2013, to August 28, 2015.

important effects. First, transmission from the EFFR to the otherunsecured rates clearly improved: the sum of lagged EFFR termsincreased from 0.23 to 0.29 in the case of the EDR and from 0.16and 0.33 in the case of the CPR. Second, the RPR became a sig-nificant predictor of the EFFR movements, in contrast to the pre-crisis relationship where RPR was moving in response to changesin the EFFR. Interestingly, RPR became highly significant in theEDR and CPR equations, emphasizing the growing importance ofthe repo market. Hence, it appears that the ON RRP markedlyimproved the overall co-movement of overnight interest rates and


Figure 6. Day of Maintenance Period Effects on EFFRduring the Pre-crisis Period

Notes: Dots indicate point estimates and horizontal lines mark the boundariesof the 95 percent confidence bands. M, T, W, R, and F denote days of the weekfrom Monday to Friday. The subscripts indicate whether the corresponding dateis the first or the second one in the maintenance period.

transmission from the federal funds market to other segments ofunsecured funding markets.

5.2 Reserve Maintenance Period Effects

In figure 6, we report point and interval estimates for the coeffi-cients of the effects of reserve maintenance days on the EFFR in thepre-crisis period. Different days of the week had small but econom-ically meaningful and statistically significant effects on the EFFR.For example, due to elevated payment flows following weekends, theEFFR used to be firmer by 1 to 2 basis points on Mondays. Bycontrast, funds used to trade softer by a slightly greater magnitudeon Fridays, as banks generally tried to avoid an excess position overthe weekend, during which reserves count for three days toward thereserve requirement. Tuesdays were also associated with softness dueto reduced demand towards the middle of the week when paymentflows are relatively lighter. These estimates are consistent with those


of Hamilton (1996), Carpenter and Demiralp (2006), and Judson andKlee (2010) that were obtained in different empirical frameworks.

In the ELB period, although we cannot statistically reject day-of-the-week effects in the federal funds market, our estimates (notreported) imply miniscule effects. When we combine our coeffi-cient estimates with trading volumes reported by Afonso, Entz, andLeSueur (2013b), we find that the average day-of-week effect is aboutonly about 3 percent of its pre-crisis level in dollar terms. Moreover,when we normalize the estimated effects by the standard deviationof the EFFR residuals to control for the dramatically different levelof the average EFFR across the two periods, we find that the day-of-week effect is about 70 percent weaker in the ELB period. Therefore,given the abundance of reserves and their fairly widespread distri-bution as reported by Ennis and Wolman (2015), we conclude thatreserve maintenance effects in the federal funds market diminishedsubstantially.

5.3 Market Dynamics on Financial Reporting Days

The estimated magnitudes of calendar effects are quite differentacross the pre-crisis and ELB periods, as is evident from panel Bin tables 4 and 5. However, the average levels of overnight inter-est rates are also dramatically different across the two samples. Tocontrol for the general level of interest rates and allow for a directcomparison between the two periods, we normalize the estimates rel-ative to standard deviations of model residuals from the respectiveequation in the VAR system.

Figure 7 shows the normalized estimates for the two main sam-ples. In the pre-crisis sample, all rates were subject to modest upwardpressure at month-ends, possibly due to heavier payment flows aswell as adjustments to balance sheets related to financial reporting.Most comprehensive financial reports are produced on a quarterlybasis, so deleveraging by financial intermediaries on quarter-end isquite common. Indeed, quarter-end effects were more prominentthan month-end effects, with the exception of the EFFR. Rates weremarkedly softer in the repo market, likely because of lower financingdemand from dealers managing their leverage. In contrast, it appearsthat reduced willingness to lend in unsecured markets on quarter-ends was the dominant factor leading to higher rates on financial


Figure 7. Month- and Quarter-End Effects

Notes: Dots indicate point estimates and horizontal lines mark the boundariesof the 95 percent confidence bands. M and Q denote month-end and quarter-end,respectively. Effects are normalized with respect to the standard deviations ofmodel residuals.

reporting days. This pattern is observed especially for LIBOR, likelyreflecting banks’ desire to show strong liquidity positions on theirfinancial statements and regulatory filings.

Money market dynamics on financial reporting days changedmaterially in the ELB sample. First of all, both the EFFR andthe EDR started softening on quarter-ends, reflecting fewer IOR-arbitrage trades by foreign banks and large domestic banks. Bal-ance sheet constraints associated with the new FDIC assessmentscheme and Basel III leverage ratio that became prevalent in thelater part of the ELB sample largely explain the reduced borrowingdemand on quarter-ends. Contrary to the case of unsecured rates,the quarter-end effect became insignificant in the repo market, onnet, likely reflecting a combination of factors. First, earlier in theELB period, collateral demand was relatively strong due to flight-to-quality flows, leading to increased willingness to lend cash at lowerrates in lieu of Treasury collateral. Second, later in the period, as newregulations were announced and implemented, lower dealer lever-age and reduced net repo financing reduced the scope of quarter-end deleveraging effects. Finally, the availability of the ON RRPas a viable investment, especially on financial reporting dates whenother investment options may be limited, reduced the potential forsharp falls in the repo rates. In terms of the CPR, cash lenders’search for alternative investments on quarter-ends amid weaker


Figure 8. Month- and Quarter-End Effects withinthe ELB Period


demand by bank borrowers appears to have led to a softening in thisrate.

Figure 8 shows the normalized calendar effects on rates throughthe ELB period. Both the EFFR and the EDR started to declinenotably at month-ends later in the ELB sample as foreign bankswithdrew from the market for financial reporting purposes. More-over, downward pressure on these rates at quarter-ends also becamemore pronounced, especially for the EDR. This likely reflects thefact that Eurodollars are a relatively more important source of dol-lar funding for foreign banks, which are subject to less stringentimplementation of the Basel III leverage ratio. In contrast, quarter-end effects on CPR have been relatively stable, suggesting limitedspillover effects from the federal funds and Eurodollar markets.The absence of direct implications on the non-financial commer-cial paper market also suggests that the leverage ratio requirementshave indeed been the primary driver for unsecured rates on financialreporting days.

5.4 Volatility and Correlation of Overnight Interest Rates

We now focus on both general and financial-reporting-driven volatil-ity dynamics across the two main sample periods as well as before


Table 8. Volatility of Rates

Pre-crisis ELB

EFFR RPR LIBOR CPR EFFR RPR EDR CPR

σε 5.64 5.91 3.94 4.30 1.05 2.38 1.22 1.51τ 0.116 0.306 0.450 0.316 0.253 0.231 0.440 0.414

(0.01) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00)δ 0.839 0.173 0.180 0.171 0.396 0.385 0.240 0.282

(0.00) (0.11) (0.03) (0.01) (0.05) (0.00) (0.00) (0.00)

Notes: p-values based on robust standard errors are reported in parentheses. σε arereported in basis points.

and after the inception of the ON RRP facility. The parameter esti-mates of the volatility models for the pre-crisis and ELB samples areshown in table 8. As expected, volatility of all rates declined sub-stantially at the ELB in absolute terms. For example, the volatilityof innovations in the EFFR equation declined from 5.6 basis pointsto only about 1 basis point. Meanwhile, the volatility process forthe EFFR has become notably less persistent, as captured by thedecline in the sum of GARCH parameters (τ + δ), and more respon-sive to shocks, as measured by the increase in the coefficient of thesquared innovation (τ). Therefore, aside from calendar effects, whichwill be discussed in more detail below, volatility clustering in theEFFR became prevalent amid subdued trading activity in the federalfunds market. Meanwhile, the volatility of the repo rate has becomesomewhat more persistent, with a slight increase in sensitivity toshocks.

Figure 9 shows the estimated month-end and quarter-end effectson volatilities in the pre-crisis and the ELB samples.23 As before,estimates are normalized by dividing by the standard deviationsof residuals to allow for direct comparison across the two periods.Prior to the crisis, similar to the calendar effects in the conditionalmean models, quarter-ends had larger effects on rate volatilities

23The confidence bands that are based on asymptotic normal distribution areasymmetric, as they are estimated in the variance space and then converted tostandard deviations.


Figure 9. Month- and Quarter-End Effects on Volatility


Table 9. Volatility of Rates within the ELB Period

Before ON RRP After ON RRP

EFFR RPR EDR CPR EFFR RPR EDR CPR

σε 1.11 2.65 1.27 1.67 0.73 1.40 0.79 0.83τ 0.212 0.159 0.365 0.383 0.189 0.315 0.458 0.158

(0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.00) (0.02)δ 0.561 0.327 0.368 0.281 0.191 0.465 0.146 0.681

(0.00) (0.00) (0.00) (0.00) (0.02) (0.01) (0.00) (0.00)

Notes: p-values based on robust standard errors are reported in parentheses. σε arereported in basis points.

than month-ends. Most notably, RPR exhibited substantial volatil-ity clustering with two to five times higher volatility on quarter-endsthan other times. In the ELB period, quarter-end volatility in theRPR moderated significantly, while becoming higher in the case ofthe EFFR.

Consistent with the soft floor set by the ON RRP facility, volatil-ity of the overnight interest rates declined 35 to 50 percent in thesecond half of the ELB sample, as shown in table 9. Moreover, theestimated parameters indicate a substantial reduction in volatilityclustering of the RPR, mainly led by a dramatic decline in the calen-dar effects (table 9 and figure 10). Figure 11 illustrates this striking


Figure 10. Month- and Quarter-End Effects on Volatilitywith the ELB Period


Figure 11. Repo Rate Volatility and ON RRP

change in the volatility of RPR following the inception of the ONRRP.24 Elevated-volatility episodes in the fall of 2013 are relatedto the Treasury’s debt limit impasse and government shutdown. Wealso observe that calendar effects on RPR volatility largely dimin-ished, likely reflecting the fact that the facility rate limiting the

24One caveat in these estimates is that the unconditional variances of theGARCH specifications are anchored to the corresponding sample variances.Therefore, the level shift after the ON RRP inception reflects the average effectacross the two ELB subsamples.


Table 10. Correlations of VAR Residuals

Pre-crisis ELB

RPR LIBOR CPR RPR EDR CPR

Normal Times 0.490 0.586 0.614 0.457 0.545 0.373(0.00) (0.00) (0.00) (0.00) (0.00) (0.00)

Month-End 0.421 0.246 0.341 0.301 0.879 0.395(0.04) (0.47) (0.22) (0.19) (0.00) (0.37)

Quarter-End 0.348 0.334 0.362 −0.056 0.564 0.360(0.30) (0.23) (0.32) (1.00) (0.03) (0.29)

Notes: Correlations with EFFR. p-values based on robust standard errors arereported in parentheses.

potential for sharp falls in rates on financial reporting dates wheninvestment options are limited for cash lenders in the repo market.

Similar to the case of the RPR, the quarter-end effect on theCPR also became insignificant in the second half of the ELB sam-ple. In contrast, month-end and quarter-end effects became morepronounced for other unsecured rates, mainly due to the pronouncedpullback from the federal funds and Eurodollar markets driven bythe Basel III regulations.

Correlation structure of VAR innovations provides furtherinsights into the co-movement of overnight interest rates. Table 10reports estimates for the pre-crisis and ELB samples from the mul-tivariate GARCH framework. Interestingly, the correlations of theEFFR residuals with those of the other three rates during normaltimes are fairly close across the two main samples. Hence, it appearsthat factors exogenous to the dynamic system of these rates, suchas Treasury debt issuance and related liquidity effects, continuedto operate in a similar fashion, on net. We also compare correla-tions within the ELB sample and report the results in table 11.Although reliable comparisons are not possible in some casesgiven the limited number of month-end and quarter-end observa-tions in the ELB subsamples, it appears that the correlation ofEFFR and RRP innovations declined over time. Meanwhile, theEFFR innovations became more strongly correlated with thoseof the EDR. This increased correlation is more pronounced on


Table 11. Correlations of VAR Residualswithin the ELB Period

Before ON RRP After ON RRP

RPR EDR CPR RPR EDR CPR

Normal Times 0.502 0.546 0.413 0.128 0.612 0.173(0.00) (0.00) (0.00) (0.16) (0.00) (0.07)

Month-End 0.395 0.596 0.104 −0.291 0.854 0.039(0.17) (0.05) (0.63) (1.00) (0.00) (0.90)

Quarter-End 0.032 0.595 0.358 −0.489 0.356 0.334(0.95) (0.05) (0.34) (1.00) (0.59) (0.51)

Notes: Correlations with EFFR. p-values based on robust standard errors arereported in parentheses.

month-ends, likely reflecting the effects of the Basel III leverage ratiorequirements.

6. Conclusion

We analyze changing dynamics of overnight money markets in thewake of the global financial crisis and the resulting policy response.To that end, we model long-run and short-run dynamics of a setof key money market rates as well as their interrelations in a mul-tivariate framework. We analyze the effects of changing monetaryand regulatory policy that evolved in the aftermath of the crisis onmoney markets.

We find that co-movement of money market rates weakened inthe ELB period compared with the historical norms. Although thefederal funds rate continued to provide an anchor for unsecuredovernight interest rates, its transmission to the repo market washampered. Moreover, the day-of-week effects on the federal fundsrate have substantially diminished, reflecting the abundance of bankreserves and their fairly widespread distribution.

New banking regulations and the Fed’s ON RRP facilityintroduced in 2013 have further transformed the money markets.Movements in unsecured short-term funding costs around financialreporting days have been exacerbated by increased costs of large


balance sheets in the new regulatory environment. Consistent withthe intended effect of the ON RRP to set a soft floor for overnightfunding rates, interest rate co-movement improved and rate volatil-ities, especially in the repo market, substantially declined after theON RRP operations started. Moreover, calendar effects in the repomarket largely disappeared, reflecting diminished potential for dropsin rates, as well as the availability of reverse repos with the Fed as aviable investment option around financial reporting days when otheralternatives are limited.

References

Adrian, T., M. Fleming, J. Goldberg, M. Lewis, F. M. Natalucci,and J. J. Wu. 2013. “Dealer Balance Sheet Capacity and Mar-ket Liquidity During the 2013 Selloff in Fixed Income Markets.”FEDS Notes (October 16). Board of Governors of the FederalReserve System.

Adrian, T., and H. S. Shin. 2010. “Liquidity and Leverage.” Journalof Financial Intermediation 19 (3): 418–37.

Afonso, G., A. Entz, and E. LeSueur. 2013a. “Who’s Borrowing inthe Fed Funds Market.” Liberty Street Economics Blog, Decem-ber 9. Available at http://libertystreeteconomics.newyorkfed.org/2013/12/whos-borrowing-in-the-fed-funds-market.html.

———. 2013b. “Who’s Lending in the Fed Funds Mar-ket.” Liberty Street Economics Blog, December 2. Availableat http://libertystreeteconomics.newyorkfed.org/2013/12/whos-lending-in-the-fed-funds-market.html.

Afonso, G., A. Kovner, and A. Schoar. 2011. “Stressed, Not Frozen:The Federal Funds Market in the Financial Crisis.” Journal ofFinance 66 (4): 1109–39.

Anbil, S., and Z. Senyuz. 2016. “Window-Dressing and TradingDynamics in the Triparty Repo Market.” SSRN Working Paper.

Anderson, A. G., J. E. Ihrig, G. C. Weinbach, and E. E. Meade.2016. “What Happened in Money Markets After the Fed’sDecember Rate Increase?” FEDS Notes (February 22). Boardof Governors of the Federal Reserve System. Available athttp://dx.doi.org/10.17016/2380-7172.1713.


Ashcraft, A. B., M. L. Bech, and W. S. Frame. 2010. “The FederalHome Loan Bank System: The Lender of Next-to-Last Resort?”Journal of Money, Credit and Banking 42 (4): 551–83.

Baklanova, V., C. Caglio, M. Cipriani, and A. Copeland. 2016. “ANew Survey of the U.S. Bilateral Repo Market: A Snapshot ofBroker-Dealer Activity.” Staff Report No. 758, Federal ReserveBank of New York.

Bech, M., E. Klee, and V. Stebunovs. 2014. “Arbitrage, Liquidityand Exit: The Repo and Federal Funds Markets Before, Dur-ing, and Emerging from the Financial Crisis.” In Developmentsin Macro-Finance Yield Curve Modelling, ed. J. S. Chadha,A. C. J. Durre, M. A. S. Joyce, and L. Sarno, 293–325 (chapter11). Cambridge: Cambridge University Press.

Bollerslev, T. 1990. “Modelling the Coherence in Short-run Nomi-nal Exchange Rates: A Multivariate Generalized ARCH Model.”Review of Economics and Statistics 72 (3): 498–505.

Carpenter, S., and S. Demiralp. 2006. “The Liquidity Effect in theFederal Funds Market: Evidence from Daily Open Market Opera-tions.” Journal of Money, Credit and Banking 38 (June): 901–20.

Choi, D. B., and H.-S. Choi. 2016. “The Effect of Monetary Pol-icy on Bank Wholesale Funding.” Staff Report No. 759, FederalReserve Bank of New York.

Cipriani, M., and J. Gouny. 2015. “The Eurodollar Market in theUnited States.” Liberty Street Economics Blog, May 27. Avail-able at http://libertystreeteconomics.newyorkfed.org/2015/05/the-eurodollar-market-in-the-united-states.html#.V6OE7KIsC1w.

Copeland, A., I. Davis, E. LeSueur, and A. Martin. 2014. “Liftingthe Veil on the U.S. Bilateral Repo Market.” Liberty Street Eco-nomics Blog, July 9. Available at http://libertystreeteconomics.newyorkfed.org/2014/07/lifting-the-veil-on-the-us-bilateral-repo-market.html#.VzowDuTYgTV.

Copeland, A., A. Martin, and M. Walker. 2014. “Repo Runs: Evi-dence from the Tri-Party Repo Market.” Journal of Finance 69(6): 2343–80.

D’Amico, S., W. English, D. Lopez-Salido, and E. Nelson. 2012.“The Federal Reserve’s Large-scale Asset Purchase Programmes:Rationale and Effects.” Economic Journal 122 (564): F415–F446.


Dickey, D. A., and W. A. Fuller. 1979. “Distribution of the Estima-tors for Autoregressive Time Series with a Unit Root.” Journalof the American Statistical Association 74 (366a): 427–31.

Elliott, G., T. J. Rothenberg, and J. H. Stock. 1996. “Efficient Testsfor an Autoregressive Unit Root.” Econometrica 64 (4): 813–36.

Engle, R. 2002. “Dynamic Conditional Correlation: A Simple Classof Multivariate Generalized Autoregressive Conditional Het-eroskedasticity Models.” Journal of Business & Economic Sta-tistics 20 (3): 339–50.

Ennis, H. M., and A. L. Wolman. 2015. “Large Excess Reserves inthe United States: A View from the Cross-Section of Banks.”International Journal of Central Banking 11 (1, January):251–87.

Frost, J., L. Logan, A. Martin, P. McCabe, F. Natalucci, and J.Remache. 2015. “Overnight RRP Operations as a Monetary Pol-icy Tool: Some Design Considerations.” Finance and Econom-ics Discussion Series No. 2015-010, Board of Governors of theFederal Reserve System.

Gorton, G., and A. Metrick. 2012. “Securitized Banking and the Runon Repo.” Journal of Financial Economics 104 (3): 425–51.

Griffiths, M. D., and D. B. Winters. 1995. “Day-of-the-Week Effectsin Federal Funds Rates: Further Empirical Findings.” Journal ofBanking & Finance 19 (7): 1265–84.

Hamilton, J. D. 1996. “The Daily Market for Federal Funds.” Jour-nal of Political Economy 104 (1): 26–56.

Hou, D., and D. Skeie. 2014. “LIBOR: Origins, Economics, Crisis,Scandal, and Reform.” Staff Report No. 667, Federal ReserveBank of New York.

Ihrig, J. E., G. C. Weinbach, and E. E. Meade. 2015. “RewritingMonetary Policy 101: What’s the Fed’s Preferred Post-CrisisApproach to Raising Interest Rates?” Journal of Economic Per-spectives 29 (4): 177–98.

Johansen, S. 1995. Likelihood-Based Inference in Cointegrated VectorAutoregressive Models. Oxford University Press.

Judson, R., and E. Klee. 2010. “Whither the Liquidity Effect: TheImpact of Federal Reserve Open Market Operations in RecentYears.” Journal of Macroeconomics 32 (3): 713–31.


Kreicher, L. L., R. N. McCauley, and P. McGuire. 2013. “The 2011FDIC Assessment on Banks’ Managed Liabilities: Interest Rateand Balance Sheet Responses.” BIS Working Paper No. 413.

Krishnamurthy, A., and A. Vissing-Jorgensen. 2011. “The Effectsof Quantitative Easing on Interest Rates: Channels and Implica-tions for Policy.” Brookings Papers on Economic Activity 2011(Fall): 215–87.

Kroeger, A., and A. Sarkar. 2016. “Monetary Policy TransmissionBefore and After the Crisis.” Liberty Street Economics Blog,June 16. Available at http://libertystreeteconomics.newyorkfed.org/2016/06/monetary-policy-transmission-before-and-after-the-crisis.html#.V6OF3KIsC1w.

Munyan, B. 2015. “Regulatory Arbitrage in Repo Markets.” Work-ing Paper No. 15-22, Office of Financial Research.

Potter, S. 2015. “Money Markets and Monetary Policy Normaliza-tion.” Remarks at the Money Marketeers of New York University,New York City, April 15. Available at https://www.newyorkfed.org/newsevents/speeches/2015/pot150415.html.

Spindt, P. A., and J. R. Hoffmeister. 1988. “The Micromechanicsof the Federal Funds Market: Implications for Day-of-the-WeekEffects in Funds Rate Variability.” Journal of Financial andQuantitative Analysis 23 (4): 401–16.

Yoldas, E., and Z. Senyuz. 2015. “Financial Stress and EquilibriumDynamics in Money Markets.” Finance and Economics Discus-sion Series No. 2015-091, Board of Governors of the FederalReserve System.

Effects of Changing Monetary and Regulatory Policy on ...

Documents

Transcript of Effects of Changing Monetary and Regulatory Policy on ...